[PATCH] perf report: Fix regression when decoding intelPT traces

[PATCH] perf report: Fix regression when decoding intelPT traces

[Mathieu Poirier]

Commit (93d10af26bb7 perf tools: Optimize sample parsing for ordered
events) breaks intelPT trace decoding by invariably returning an error if
the event type isn't a PERF_SAMPLE_TIME.

With this patch the timestamp is initialised and processing is allowed to
continue even if an error is returned by function
perf_evlist__parse_sample_timestamp().

Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
 tools/perf/util/session.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 54e30f1bcbd7..20cdcf14232b 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -1512,7 +1512,7 @@ static s64 perf_session__process_event(struct perf_session *session,
 
 		ret = perf_evlist__parse_sample_timestamp(evlist, event, &timestamp);
 		if (ret)
-			return ret;
+			timestamp = -1ULL;
 
 		ret = perf_session__queue_event(session, event, timestamp, file_offset);
 		if (ret != -ETIME)
-- 
2.7.4

Re: [PATCH] perf report: Fix regression when decoding intelPT traces

[Jiri Olsa]

On Wed, Dec 13, 2017 at 05:57:54PM -0700, Mathieu Poirier wrote:
> Commit (93d10af26bb7 perf tools: Optimize sample parsing for ordered
> events) breaks intelPT trace decoding by invariably returning an error if
> the event type isn't a PERF_SAMPLE_TIME.

right, thanks for catchng that

> 
> With this patch the timestamp is initialised and processing is allowed to
> continue even if an error is returned by function
> perf_evlist__parse_sample_timestamp().
> 
> Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
> ---
>  tools/perf/util/session.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
> index 54e30f1bcbd7..20cdcf14232b 100644
> --- a/tools/perf/util/session.c
> +++ b/tools/perf/util/session.c
> @@ -1512,7 +1512,7 @@ static s64 perf_session__process_event(struct perf_session *session,
>  
>  		ret = perf_evlist__parse_sample_timestamp(evlist, event, &timestamp);
>  		if (ret)
> -			return ret;
> +			timestamp = -1ULL;

we still have some -EFAULT error codes in there that we
want to catch.. should it be more like in below?

thanks,
jirka


---
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 54e30f1bcbd7..9498aa0efe39 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -1508,10 +1508,10 @@ static s64 perf_session__process_event(struct perf_session *session,
 		return perf_session__process_user_event(session, event, file_offset);
 
 	if (tool->ordered_events) {
-		u64 timestamp;
+		u64 timestamp = -1ULL
 
 		ret = perf_evlist__parse_sample_timestamp(evlist, event, &timestamp);
-		if (ret)
+		if (ret != -1)
 			return ret;
 
 		ret = perf_session__queue_event(session, event, timestamp, file_offset);

Re: [PATCH] perf report: Fix regression when decoding intelPT traces

[Mathieu Poirier]

On 14 December 2017 at 08:28, Jiri Olsa <jolsa@redhat.com> wrote:
> On Wed, Dec 13, 2017 at 05:57:54PM -0700, Mathieu Poirier wrote:
>> Commit (93d10af26bb7 perf tools: Optimize sample parsing for ordered
>> events) breaks intelPT trace decoding by invariably returning an error if
>> the event type isn't a PERF_SAMPLE_TIME.
>
> right, thanks for catchng that
>
>>
>> With this patch the timestamp is initialised and processing is allowed to
>> continue even if an error is returned by function
>> perf_evlist__parse_sample_timestamp().
>>
>> Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
>> ---
>>  tools/perf/util/session.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
>> index 54e30f1bcbd7..20cdcf14232b 100644
>> --- a/tools/perf/util/session.c
>> +++ b/tools/perf/util/session.c
>> @@ -1512,7 +1512,7 @@ static s64 perf_session__process_event(struct perf_session *session,
>>
>>               ret = perf_evlist__parse_sample_timestamp(evlist, event, &timestamp);
>>               if (ret)
>> -                     return ret;
>> +                     timestamp = -1ULL;
>
> we still have some -EFAULT error codes in there that we
> want to catch.. should it be more like in below?
>
> thanks,
> jirka
>
>
> ---
> diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
> index 54e30f1bcbd7..9498aa0efe39 100644
> --- a/tools/perf/util/session.c
> +++ b/tools/perf/util/session.c
> @@ -1508,10 +1508,10 @@ static s64 perf_session__process_event(struct perf_session *session,
>                 return perf_session__process_user_event(session, event, file_offset);
>
>         if (tool->ordered_events) {
> -               u64 timestamp;
> +               u64 timestamp = -1ULL
>
>                 ret = perf_evlist__parse_sample_timestamp(evlist, event, &timestamp);
> -               if (ret)
> +               if (ret != -1)
>                         return ret;

Yes, that's better  - I'll send another revision.

>
>                 ret = perf_session__queue_event(session, event, timestamp, file_offset);